Your assistant is watching. Who watches it?

Map the people, systems, and missing controls around an AI assistant. Your evidence stays in this browser until you export it.

Build your evidence map

What can it observe?
Who can review the data?
Use documents or settings you can verify. This tool does not inspect your account.

Turn permissions into a finding.

The score is an evidence organizer, not a certification. Each point is visible so you can challenge the result and verify the controls that matter.

Complete the audit to reveal the watchers, exposure, and control gaps.

A disclosed model. No hidden verdict.

0–19

Every point has a reason

Observed data adds up to six points. Watchers add up to four. Retention, sharing, access logs, and oversight contribute the remaining nine.

Local

Your evidence stays here

The record uses browser storage on this device. Export is initiated by you, and nothing is sent to a server by the page.

0–4: Limited

Access and review appear narrow based on the entered evidence. Verify that controls remain effective.

5–9: Guarded

Some exposure or control uncertainty deserves review before broader use.

10–14: Elevated
15–19: Critical

Broad observation, additional watchers, or weak accountability create material surveillance exposure.

Questions worth taking to the provider.

Collection

Which permissions are necessary for the feature I use, and can each one be disabled independently?

Verify scope

Review

Which employees, contractors, or vendors can access my content, under what trigger, and with what approval?

Name watchers

Retention

When is content deleted from primary systems and backups, and can I initiate that deletion?

Set a limit

Accountability

Can I inspect access events, contest misuse, and reach an independent reviewer?

Close the loop

Mapping AI Surveillance Exposure

Read the explanation

The audit assesses AI assistant surveillance by aggregating observation scopes and reviewer access into a nineteen-point model. Six data streams contribute up to six points, while external reviewers add up to four points across providers, employers, and contractors. Retention schedules, model training usage, and absent access logs add up to nine additional points, elevating the risk into critical status. Clicking Calculate Exposure summarizes verified oversight gaps into targeted provider questions and enables downloading a local JSON audit report.

Super generates helpful tools and automates fact-checking across the internet proactively. If you enjoyed this tool, build your own with Super and share it with a friend.